Winning percentage measures regular-season success as wins divided by total regular-season decisions: wins / (wins + losses). The resulting value ranges as a proportion, so 0.50 separates teams with winning records from teams with losing records. section 12.7.1 section 12.8.1
Use it when raw win totals alone do not make comparisons clear. A percentage can summarize how often a game condition led to a final win—for example, after a team won a particular quarter—and can also group team-seasons for statistical comparison. section 6.4.1 section 12.8.1
A practical R workflow is:
- Keep regular-season win and loss counts, then derive pct as wins divided by wins plus losses. section 12.7.1
- For a condition such as winning the first quarter, count each final-result combination and divide each count by the number of eligible games. section 6.4.1
- Plot counts alongside percentages: the measures give complementary views, while labels and ordering make the comparison easier to read. section 6.4.1
- Group teams into percentage bands or above/below 0.50 groups before comparing another measure such as a Gini coefficient. section 12.8.1 section 12.9
Winning the first quarter
After excluding first-quarter ties, overtime games, and Orlando games from the 2018–19 and 2019–20 regular seasons, teams that won the first quarter also won the game 66.98% of the time (783 home-team cases plus 550 road-team cases out of 1,990 eligible games). Home teams won about 73% of such games, compared with about 60% for road teams. section 6.4.1
count(nbadf3, Q1vF) %>% arrange(desc(n)) -> tbl1 tbl1 %>% mutate(pct_total = n/nrow(nbadf3)*100) %>% mutate(cum_n = cumsum(n)) %>% mutate(cum_pct_total = cumsum(pct_total)) -> tbl1 tbl1$pct_total <- round(tbl1$pct_total, digits = 2) tbl1$cum_pct_total <- round(tbl1$cum_pct_total, digits = 2) print(tbl1) ## Q1vF n pct_total cum_n cum_pct_total ## 1 HH 783 39.35 783 39.35 ## 2 RR 550 27.64 1333 66.98 ## 3 RH 365 18.34 1698 85.33 ## 4 HR 292 14.67 1990 100.00
This calculation turns counts of quarter-result and final-result combinations into percentages of all eligible games, then rounds the displayed values. section 6.4.1
Winning percentage can support association studies, not just standings. Across NBA team-seasons, teams in higher salary-inequality bands had higher average regular-season winning percentages in the displayed summaries; comparisons of winning and losing teams were also tested with a t-test. section 12.9 section 12.8.1
